Front-door Adjustment Sets

This repository contains the code for the paper "Finding and Listing Front-door Adjustment Sets" by Hyunchai Jeong, Jin Tian, and Elias Bareinboim.

How to install

Please run the following commands to install the package:

git clone https://github.com/CausalAILab/FrontdoorAdjustmentSets.git
cd FrontdoorAdjustmentSets
pip install -r requirements.txt
pip install -e .

How to run

To run some examples, try running the following commands:

python3 main.py find graphs/fig1a.txt

python3 main.py list graphs/fig1b.txt

1. First argument: find or list. find is used for finding any valid front-door adjustment set. Specify list for enumerating all valid front-door adjustment sets.
2. Second argument: graphs/fig1a.txt. the path to a file that contains all necessary information, such as graph, query, and constraints (i.e., I and R). Please check the formatting for details.

Formatting of the input file

Consider graphs/fig1b.txt as an example.

<NODES>

Each line represents the name of a variable/node (e.g., X and A).

<EDGES>

Each line describes an edge. Two types of edges are supported:

1. Directed edge: X -> A represents a directed edge from X to A.
2. Bidirected edge: X -- Y means there exists some unmeasured confounder (i.e., latent variable) between X and Y.

<TASK>

The task represents the query of interest. The lines:

treatment: X
outcome: Y

specifies a casual effect from X to Y. Multiple nodes may be specified by separating each name of a node with comma (e.g., treatment: A, B, C).

<CONSTRAINTS>

The lines:

I: B
R: A,B,C,D

represent the constraints as follows:

1. I specifies a set of nodes that must be in any candidate front-door adjustment set. If I is not specifed (i.e., by removing the line I: B), then I will be an empty set by default.
2. R specifies a set of nodes that could be in any candidate front-door adjustment set. If R is not specifed (i.e., by removing the line R: A,B,C,D), then every node but treatment and outcome nodes will be included by default.